Claude Opus 4.8 released
Anthropic · Model release
Source confirmed
Data as of October 10, 2026 · How the score is built
Superseded.Anthropic has newer models in this line:Claude Opus 5
SupersededProprietaryReasoning
Claude Opus 4.8
Decision readingClaude Opus 4.8 scores 69.1 out of 100 and ranks #18 of 218. This profile shows 40 source-displayable benchmark rows; its strongest eligible category is Mathematics at #2. API pricing is $5 input and $25 output per million tokens.
Released May 28, 2026 — see all recent releases
Claude Opus 4.8 will be repriced, updated or retired. Get each notice with its source and date. Follow model changes
Decision snapshot
Each value carries a field reference instead of floating alone. Markers compare this model with the current ranked and priced catalog; they are not absolute quality thresholds.
Capability
69.1/100
field median 50.1#18 of 218 ranked models
Public
#18of 218
Verified #13 of 83
Price
$5input / $25 output
input median $0.95blended $15
Speed
64tok/s
field median 97 tok/sFirst token 52.1 s
Context
1Mtokens
field median 256,000Maximum output length is tracked separately
Strongest published evidence
Mathematics ranks #2. Particularly strong for mathematical reasoning, scientific computing, and quantitative analysis.
Validate before choosing
40 published rows leave some tracked benchmark slots empty. Reasoning is its lowest eligible category at #24.
Source-linked · 40 displayable benchmark rows
Follow model changesCheck the published record for the model and serving route you use.
Check this modelCategory score record
Scores and ranks appear only where published evidence can be displayed. The table keeps the score, weight, cohort, and evidence state together.
| Category | Score | Rank | Percentile | Weight | Benchmarks | Evidence |
|---|---|---|---|---|---|---|
| AgenticRank #20 of 123Percentile 84thWeight 22%12 benchmarksVerified | 61.6 | 12 benchmarks | Verified | |||
| CodingRank #19 of 146Percentile 88thWeight 20%11 benchmarksVerified | 60.8 | 11 benchmarks | Verified | |||
| ReasoningRank #24 of 28Percentile 15thWeight 17%3 benchmarksVerified | 62.9 | 3 benchmarks | Verified | |||
| MultimodalRank #5 of 54Percentile 92ndWeight 12%4 benchmarksVerified | 91.2 | 4 benchmarks | Verified | |||
| KnowledgeRank #14 of 177Percentile 93rdWeight 12%6 benchmarksVerified | 70.7 | 6 benchmarks | Verified | |||
| MultilingualWeight 7%1 benchmarkVerified | Score pending | 1 benchmark | Verified | |||
| Inst. FollowingWeight 5%0 benchmarksNot measured | Not measured | 0 benchmarks | Not measured | |||
| MathRank #2 of 7Percentile 83rdWeight 5%3 benchmarksVerified | 65.7 | 3 benchmarks | Verified |
40 of 667 tracked benchmark slots have displayable evidence · bars run 0–100
Coverage detailsHow much of this is verified
Coverage is split by category so a strong number never hides a thin evidence base. Verified means the row is tied to a published source; provisional rows remain visible but separate.
- Agentic12/12 verified
- Coding11/11 verified
- Reasoning3/3 verified
- Multimodal4/4 verified
- Knowledge6/6 verified
- Multilingual1/1 verified
- Inst. FollowingNot measured
- Math3/3 verified
Verified sourceProvisionalNot measured
Capability shape
Each axis shows percentile within that category’s eligible cohort. The comparison outline is the median of the six nearest public-score peers; a collapsed vertex means the category is not rank-eligible.
Claude Opus 4.8 category percentile values
- Agentic84th percentile
- Coding88th percentile
- Reasoning15th percentile
- Multimodal92nd percentile
- Knowledge93rd percentile
- MultilingualNot eligible
- Instruction followingNot eligible
- Math83rd percentile
The dashed outline is median of 6 nearest peers.
Eligible category ranks
- Agentic#20/123
- Coding#19/146
- Reasoning#24/28
- Multimodal#5/54
- Knowledge#14/177
- MultilingualNot ranked
- Inst. FollowingNot ranked
- Math#2/7
Top decileTop quartileMid-fieldNot eligible
Benchmark ledger
Coding opens by default. The marker compares each value with the best source-verified result in the catalog; provisional leaders do not set the reference. Expand the remaining categories for every published row.
Coding11 rows
| Benchmark | Score | Versus best verified row | Gap | Weight | Evidence |
|---|---|---|---|---|---|
| SWE-bench Pro | Score69.2% | Versus best verified row Best verified: Claude Opus 5.5 · 89.9% | Gap20.7 behind | Weight26% ref. weight | Provider exact |
| CursorBench 3.2 | Score62.3% | Versus best verified row Best verified: Claude Fable 5.1 · 73.4% | Gap11.1 behind | Weight10% ref. weight | Benchmark exact |
| LiveCodeBench (Vals)LiveCodeBench, Vals AI run | Score87.8% | Versus best verified row Best verified: Claude Fable 5.1 · 90.5% | Gap2.7 behind | Weight8% ref. weight | |
| SWE Multilingual | Score84.4% | Versus best verified row Best verified: Claude Opus 5.5 · 93.9% | Gap9.5 behind | Weight5% ref. weight | Provider exact |
| FrontierCode 1.1 Main | Score46.5% | Versus best verified row Best verified: Claude Opus 5.5 · 54.4% | Gap7.9 behind | Weight4% ref. weight | Benchmark exact |
| Terminal-Bench 2.1 | Score74.6% | Versus best verified row Best verified: SWE-2 · 92.8% | Gap18.2 behind | WeightScored in Agentic | Provider exact |
| SWE-bench VerifiedSoftware Engineering Benchmark Verified | Score88.6% | Versus best verified row Best verified: Claude Opus 5 · 96% | Gap7.4 behind | WeightDisplay only | Provider exact |
| SWE MultimodalSWE-bench Multimodal | Score38.4% | Versus best verified row Best verified: Claude Opus 5.5 · 61.4% | Gap23 behind | WeightDisplay only | Provider exact |
| cursorBench31 | Score58.4% | Versus best verified row Best verified: Claude Fable 5 · 70.6% | Gap12.2 behind | WeightDisplay only | Benchmark exact |
| SWE-bench (Vals)SWE-bench, Vals AI run | Score88.6% | Versus best verified row Best verified: Claude Opus 5 · 97.0% | Gap8.4 behind | WeightDisplay only | Verified |
| PostTrainBench v1.1 | Score32.9% | Versus best verified row Best verified: Claude Opus 5.5 · 49.3% | Gap16.4 behind | WeightDisplay only | Benchmark exact |
Agentic12 rows
| Benchmark | Score | Versus best verified row | Gap | Weight | Evidence |
|---|---|---|---|---|---|
| OSWorld 2.0 | Score20.6% | Versus best verified row Best verified: GPT-6 Astra · 72.6% | Gap52 behind | Weight10% ref. weight | Benchmark exact |
| Terminal-Bench 2.1 | Score74.6% | Versus best verified row Best verified: SWE-2 · 92.8% | Gap18.2 behind | Weight8% ref. weight | Provider exact |
| BrowseComp | Score84.3% | Versus best verified row Best verified: Atria Dawn Preview · 92.5% | Gap8.2 behind | Weight8% ref. weight | Provider exact |
| OSWorld-Verified | Score83.4% | Versus best verified row Best verified: Qwen3.8 Max · 86.1% | Gap2.7 behind | Weight6% ref. weight | Provider exact |
| MCP Atlas | Score82.2% | Versus best verified row Best verified: Muse Spark 1.1 · 88.1% | Gap5.9 behind | Weight4% ref. weight | Provider exact |
| Terminal-Bench 2.1 (Vals)Terminal-Bench 2.1, Vals AI run | Score71.9% | Versus best verified row Best verified: GPT-6 Astra · 87.3% | Gap15.4 behind | Weight3% ref. weight | |
| Terminal-Bench 3.0 | Score21.1% | Versus best verified row Best verified: Claude Opus 5 · 42.7% | Gap21.6 behind | Weight3% ref. weight | Benchmark exact |
| Toolathlon | Score59.9% | Versus best verified row Best verified: Muse Spark 1.1 · 75.6% | Gap15.7 behind | Weight3% ref. weight | Provider exact |
| DeepSearchQA | Score93.1% | Versus best verified row Best verified: Atria Dawn Preview · 96.0% | Gap2.9 behind | Weight2% ref. weight | Provider exact |
| Finance Agent v2 | Score53.9% | Versus best verified row Best verified: Gemini 4 Argon · 65.4% | Gap11.5 behind | WeightDisplay only | Provider exact |
| Gert LabsGert Labs Composite Game Benchmark | Score72.97% | Versus best verified row Best verified: Claude Opus 4.8 · 72.97% | GapBest verified | WeightDisplay only | Benchmark exact |
| ResearchClawBench | Score21.1% | Versus best verified row Best verified: Claude Opus 4.8 · 21.1% | GapBest verified | WeightDisplay only | Benchmark exact |
Reasoning3 rows
| Benchmark | Score | Versus best verified row | Gap | Weight | Evidence |
|---|---|---|---|---|---|
| ARC-AGI-2Abstraction and Reasoning Corpus for AGI v2 | Score72.1% | Versus best verified row Best verified: GPT-6 Astra · 95% | Gap22.9 behind | WeightWeighted 25% | Benchmark exact |
| ARC-AGI-3Abstraction and Reasoning Corpus for AGI v3 | Score1.5% | Versus best verified row Best verified: GPT-6 Astra · 62.7% | Gap61.2 behind | WeightWeighted 15% | Benchmark exact |
| ARC-AGI-1ARC-AGI-1 Semi-Private Evaluation | Score92.50% | Versus best verified row Best verified: GPT-6 Astra · 98.50% | Gap6 behind | WeightDisplay only | Benchmark exact |
Multimodal4 rows
| Benchmark | Score | Versus best verified row | Gap | Weight | Evidence |
|---|---|---|---|---|---|
| OfficeQA Pro | Score66.2% | Versus best verified row Best verified: Claude Opus 5.5 · 67.7% | Gap1.5 behind | WeightWeighted 25% | Provider exact |
| CharXivCharXiv Reasoning | Score89.9% | Versus best verified row Best verified: Qwen3.8 Max · 93.5% | Gap3.6 behind | WeightWeighted 20% | Provider exact |
| ScreenSpot Pro | Score87.9% | Versus best verified row Best verified: GPT-6 Astra · 92.7% | Gap4.8 behind | WeightDisplay only | Provider exact |
| CharXiv w/o toolsCharXiv Reasoning without tools | Score80.5% | Versus best verified row Best verified: Claude Mythos 5 · 88.9% | Gap8.4 behind | WeightDisplay only | Provider exact |
Knowledge6 rows
| Benchmark | Score | Versus best verified row | Gap | Weight | Evidence |
|---|---|---|---|---|---|
| HLEHumanity's Last Exam | Score57.9% | Versus best verified row Best verified: Claude Fable 5.1 · 65% | Gap7.1 behind | Weight44% ref. weight | Provider exact |
| HLE w/o toolsHumanity's Last Exam without tools | Score49.8% | Versus best verified row Best verified: Claude Opus 5.5 · 64.4% | Gap14.6 behind | Weight7% ref. weight | Provider exact |
| MMLU-Pro (Vals)MMLU-Pro, Vals AI run | Score89.6% | Versus best verified row Best verified: Claude Fable 5.1 · 92.4% | Gap2.8 behind | Weight6% ref. weight | Verified |
| GPQAGraduate-Level Google-Proof Q&A | Score93.6% | Versus best verified row Best verified: GPT-6 Astra · 96% | Gap2.4 behind | Weight3% ref. weight | Provider exact |
| GPQA Diamond (Vals)GPQA Diamond, Vals AI run | Score92.4% | Versus best verified row Best verified: Gemini 3.1 Pro · 95.5% | Gap3.1 behind | Weight2% ref. weight | |
| GPQA-DGPQA Diamond | Score93.6% | Versus best verified row Best verified: GPT-6 Astra · 96.0% | Gap2.4 behind | WeightDisplay only | Provider exact |
Multilingual1 row
| Benchmark | Score | Versus best verified row | Gap | Weight | Evidence |
|---|---|---|---|---|---|
| INCLUDE | Score87.6% | Versus best verified row Best verified: Claude Opus 5 · 89.8% | Gap2.2 behind | WeightDisplay only | Provider exact |
Math3 rows
| Benchmark | Score | Versus best verified row | Gap | Weight | Evidence |
|---|---|---|---|---|---|
| FrontierMath v2 (Tiers 1-3)FrontierMath v2 Tiers 1-3 | Score47.241% | Versus best verified row Best verified: GPT-5.6 Sol · 89.000% | Gap41.8 behind | WeightWeighted 30% | Benchmark exact |
| USAMO 2026United States of America Mathematical Olympiad 2026 | Score96.7% | Versus best verified row Best verified: Claude Mythos 5 · 97.6% | Gap0.9 behind | WeightWeighted 10% | Provider exact |
| FrontierMath v2 (Tier 4)FrontierMath v2 Tier 4 | Score31.250% | Versus best verified row Best verified: GPT-6 Astra · 97.600% | Gap66.4 behind | WeightWeighted 10% | Benchmark exact |
Bars run 0–100; the dark tick marks the best source-verified value
All 40 rowsWhat it costs to get this score
Published API price against the public score. The x-axis uses a log scale; the dashed path marks models that are not beaten by a cheaper, higher-scoring option. Price uses average of published input and output rates.
Current modelExplore all models
Claude Opus 4.8 · 69.1 score · $15 blended per million tokens
The chart opens on the current model. Scroll horizontally to inspect the full price axis.
Horizontal: blended price per million tokens, log scale · Vertical: public score
Lineage
The sequence follows explicit supersedes links. Each score is estimated for that model; a relative can inform a sparse estimate but never sets a floor, so a newer release can score below an earlier one. Scores and prices remain blank when the corresponding public row or first-party rate is unavailable.
Base entry
Radar
Full release historyClaude Opus 4.8 release history
Radar confirmed these at the source. Use Claude Opus 4.8 in your work? Explore Radar to follow supported changes and choose your alerts.
Radar
Spec sheet
Each documented value carries its source. Missing fields stay visible as not sourced or not published, rather than disappearing from the page.
- API model ID
- claude-opus-4-8Anthropic model overview
- Context window
- 1MAnthropic model overview
- Maximum output
- Not sourced yet
- Knowledge cutoff
- Not sourced yet
- Input modalities
- text, imageAnthropic model overview
- Output modalities
- textAnthropic model overview
- Parameters
- Not disclosed by the provider
- Availability
- Claude APIAnthropic model overview
- Cloud regions
- Not tracked yet
- Lifecycle
- activeAnthropic model overview
- API capabilities
- Tool calling, structured outputs, and batch support are not tracked yet
- Prompt caching
- Not documented in the pricing record
- Self-host
- Weights are not published
- Rate limits
- Not tracked yet
How to read this profile
The visual layer above carries the decisions. These notes preserve the model, ranking, coverage, and family context behind the numbers.
Claude Opus 4.8 ranks #18 of 218 on the public leaderboard with a score of 69.12/100. Its source-verified position is #13 of 83.
Claude Opus 4.8 is a proprietary model with a 1M context window. It uses an explicit reasoning mode, which can improve complex problem solving while adding latency and token use.
Anthropic's May 28, 2026 Claude Opus 4.8 release, supplemented by CursorBench v3.1 and CursorBench v3.2's Opus 4.8 Max setting. BenchLM maps the official system-card capability snapshot using adaptive thinking / max-effort settings where Anthropic reports them, and keeps PDF-only metrics with incompatible or newly introduced scoring surfaces in the Opus 4.8 benchmark backlog note.
Its explicit predecessor is Claude Opus 4.7 (Adaptive). 40 of 667 tracked benchmark slots currently have displayable evidence. Missing categories stay blank.
Its strongest eligible category is Mathematics at #2, while its lowest eligible position is Reasoning at #24. particularly strong for mathematical reasoning, scientific computing, and quantitative analysis.
Last updated October 10, 2026. Runtime fields remain blank until a sourced snapshot exists.
Questions
How does Claude Opus 4.8 perform overall in AI benchmarks?
Claude Opus 4.8 ranks #18 out of 218 models on the public BenchAlign leaderboard, with a score of 69.12/100. Its evidence status is Supported, and this profile shows 40 source-displayable benchmark rows. The label describes evidence depth, not a provider quality claim; inspect category rows before choosing a workload.
Is Claude Opus 4.8 good for knowledge and understanding?
Claude Opus 4.8 ranks #14 out of 177 eligible models for knowledge and understanding, with a public category score of 70.7/100. Higher-ranked alternatives are available for workloads where this category decides the choice. Check the underlying rows before treating the aggregate as a workload guarantee.
Is Claude Opus 4.8 good for coding and programming?
Claude Opus 4.8 ranks #19 out of 146 eligible models for coding and programming, with a public category score of 60.8/100. Higher-ranked alternatives are available for workloads where this category decides the choice. Check the underlying rows before treating the aggregate as a workload guarantee.
Is Claude Opus 4.8 good for mathematics?
Claude Opus 4.8 ranks #2 out of 7 eligible models for mathematics, with a public category score of 65.7/100. That places it in the current top ten for this category. Check the underlying rows before treating the aggregate as a workload guarantee.
Is Claude Opus 4.8 good for reasoning and logic?
Claude Opus 4.8 ranks #24 out of 28 eligible models for reasoning and logic, with a public category score of 62.9/100. Higher-ranked alternatives are available for workloads where this category decides the choice. Check the underlying rows before treating the aggregate as a workload guarantee.
Is Claude Opus 4.8 good for agentic tool use and computer tasks?
Claude Opus 4.8 ranks #20 out of 123 eligible models for agentic tool use and computer tasks, with a public category score of 61.6/100. Higher-ranked alternatives are available for workloads where this category decides the choice. Check the underlying rows before treating the aggregate as a workload guarantee.
Is Claude Opus 4.8 good for multimodal and grounded tasks?
Claude Opus 4.8 ranks #5 out of 54 eligible models for multimodal and grounded tasks, with a public category score of 91.2/100. That places it in the current top ten for this category. Check the underlying rows before treating the aggregate as a workload guarantee.
Is Claude Opus 4.8 good for multilingual tasks?
Claude Opus 4.8 has source-displayable benchmark coverage for multilingual tasks, but the public category table does not assign it a rank there. The individual rows remain available for inspection. A missing category position means the evidence threshold was not met; it does not convert the model's unmeasured work into a zero.
Does Claude Opus 4.8 have full benchmark coverage on BenchLM?
No. Claude Opus 4.8 currently has 56 source-displayable rows across 667 tracked benchmark slots. The profile exposes published, non-generated evidence and leaves missing categories blank until an exact evaluation is available. Coverage describes how much was measured; it is not a penalty added to an individual benchmark result.
What is the context window size of Claude Opus 4.8?
Claude Opus 4.8 has a documented context window of 1M. That figure is the maximum combined prompt and retained-conversation space reported for this exact model; it is not the maximum output length. The profile keeps output limits separate because providers often publish those limits independently.